vs
Gemma 4 26B vs GPT-5.5: which is better at resume extraction?
GPT-5.5 leads Gemma 4 26B by 20.7 points (83.5 vs 62.8 F1). Cost per 1,000 pages: $0.56 vs $87. Median time per resume: 29 s vs 48 s.
Gemma 4 26B wins on 26 of 300 resumes, GPT-5.5 on 270, 4 tied.
62.82
Gemma 4 26B
A · Vision LLM
−20.68A − B, points of F1
83.50
GPT-5.5
B · Vision LLM
A better on 26level on 4B better on 270
GPT-5.5 is ahead by 20.68 points on these 300 resumes.
- Many entries 83A 58.0 · B 83.8−25.8
- Long bullet lists 101A 57.4 · B 84.3−26.9
- Split across pages 85A 60.0 · B 85.0−25.1
- Big skills list 60A 61.9 · B 85.5−23.6
- Custom sections 105A 58.9 · B 80.8−21.9
- Combined headings 37A 60.9 · B 83.4−22.5
- Promotions 88A 59.5 · B 84.0−24.5
- Partial dates 138A 62.9 · B 84.0−21.1
- Overlapping jobs 69A 57.5 · B 82.1−24.6
- Accented text 13A 59.3 · B 83.9−24.6
Side by side
| Gemma 4 26B | GPT-5.5 | |
|---|---|---|
| Type | Vision LLM | Vision LLM |
| Maker | OpenAI | |
| Rank | 16 of 18 | 11 of 18 |
| F1 | 62.8 | 83.5 |
| Precision | 69.1 | 84.9 |
| Recall | 66.2 | 86.0 |
| Cost per 1,000 pages | $0.56 | $87 |
| Median time per resume | 29 s | 48 s |
| Resumes won | 26 | 270 |
Where each one wins
Where the two differ most.
| Kind of resume | # of resumes | Gemma 4 26B | GPT-5.5 | Gap |
|---|---|---|---|---|
| Content as images (distractors) | 8 | 65.6 | 92.6 | −26.9 |
| Long bullet lists (extraction challenges) | 101 | 57.4 | 84.3 | −26.9 |
| References (noise) | 55 | 62.5 | 88.5 | −25.9 |
| Many entries (extraction challenges) | 83 | 58.0 | 83.8 | −25.8 |
| Resume-builder templates (source) | 10 | 58.6 | 83.9 | −25.3 |
| Split across pages (extraction challenges) | 85 | 60.0 | 85.0 | −25.1 |
| Wording traps (distractors) | 16 | 60.6 | 85.3 | −24.7 |
| Overlapping jobs (extraction challenges) | 69 | 57.5 | 82.1 | −24.6 |
| Accented text (extraction challenges) | 13 | 59.3 | 83.9 | −24.6 |
| Promotions (extraction challenges) | 88 | 59.5 | 84.0 | −24.5 |
| Distractor layouts (source) | 18 | 62.4 | 86.6 | −24.2 |
| Big skills list (extraction challenges) | 60 | 61.9 | 85.5 | −23.6 |
| Section | Resumes | Gemma 4 26B | GPT-5.5 | Gap |
|---|---|---|---|---|
| Contact details | 300 | 84.6 | 93.5 | −8.9 |
| Experience | 300 | 56.0 | 87.4 | −31.4 |
| Education | 297 | 68.0 | 88.7 | −20.8 |
| Summary | 290 | 35.5 | 81.8 | −46.3 |
| Certifications | 244 | 63.0 | 77.3 | −14.4 |
| Awards | 110 | 55.7 | 75.3 | −19.5 |
| Skills | 290 | 81.9 | 91.5 | −9.6 |
| Languages | 171 | 80.9 | 85.7 | −4.8 |
| Custom sections | 223 | 32.3 | 49.7 | −17.4 |
| Interests | 67 | 71.8 | 92.4 | −20.6 |
| Volunteering | 78 | 59.4 | 85.0 | −25.6 |
| Profiles | 37 | 57.2 | 68.5 | −11.3 |
| Projects | 83 | 55.0 | 86.4 | −31.4 |
| Publications | 53 | 54.5 | 82.5 | −28.0 |
Published resumes they disagree on most
Open one to see both outputs field by field.
| Resume | Gemma 4 26B | GPT-5.5 | Gap |
|---|---|---|---|
| Creative Director: Accented text in a styled CV | 39.3 | 90.2 | −50.9 |
| Supervisory Human Resources Specialist, GS-0201-15: Resume-builder template | 29.4 | 79.9 | −50.4 |
| Deputy Head of Mission: Rotated scan with hidden identity text | 34.1 | 64.7 | −30.6 |
| Head of Developer Experience (Internal title: Jedi Master of Analytics) : Unusual job titles | 53.7 | 80.2 | −26.5 |
Which to pick
GPT-5.5 is more accurate. Gemma 4 26B costs 99% less for 20.7 fewer points.
GPT-5.5 does better on content as images, long bullet lists and references.
Gemma 4 26B against every other model
- Gemma 4 26B vs LlamaParse Extract Agentic Plus
- Gemma 4 26B vs LlamaParse Extract Agentic
- Gemma 4 26B vs LlamaParse Extract Cost-Effective
- Gemma 4 26B vs Reducto Deep Extract
- Gemma 4 26B vs Reducto Extract
- Gemma 4 26B vs Extend Max Context
- Gemma 4 26B vs Extend Extract
- Gemma 4 26B vs GPT-5.6
- Gemma 4 26B vs GPT-5.4 Nano
- Gemma 4 26B vs Claude Opus 4.8
- Gemma 4 26B vs Gemini 3.5 Flash
- Gemma 4 26B vs Qwen 3.8 Flash
- Gemma 4 26B vs Qwen 3.8 27B
- Gemma 4 26B vs Qwen 3.5 35B
- Gemma 4 26B vs Qwen 3.5 9B
- Gemma 4 26B vs GLM 5.3 Flash
GPT-5.5 against every other model
- GPT-5.5 vs LlamaParse Extract Agentic Plus
- GPT-5.5 vs LlamaParse Extract Agentic
- GPT-5.5 vs LlamaParse Extract Cost-Effective
- GPT-5.5 vs Reducto Deep Extract
- GPT-5.5 vs Reducto Extract
- GPT-5.5 vs Extend Max Context
- GPT-5.5 vs Extend Extract
- GPT-5.5 vs GPT-5.6
- GPT-5.5 vs GPT-5.4 Nano
- GPT-5.5 vs Claude Opus 4.8
- GPT-5.5 vs Gemini 3.5 Flash
- GPT-5.5 vs Qwen 3.8 Flash
- GPT-5.5 vs Qwen 3.8 27B
- GPT-5.5 vs Qwen 3.5 35B
- GPT-5.5 vs Qwen 3.5 9B
- GPT-5.5 vs GLM 5.3 Flash
Popular comparisons
- LlamaParse Extract Agentic Plus vs Reducto Deep Extract
- Extend Max Context vs Reducto Deep Extract
- Extend Max Context vs LlamaParse Extract Agentic Plus
- Claude Opus 4.8 vs GPT-5.6
- Gemini 3.5 Flash vs GPT-5.6
- Claude Opus 4.8 vs Gemini 3.5 Flash
- LlamaParse Extract Agentic vs LlamaParse Extract Agentic Plus
- Reducto Deep Extract vs Reducto Extract
- Gemma 4 26B vs Qwen 3.8 27B
- Document APIs vs vision LLMs